about the evaluation script of ST-RRED

[redrock303]

Could you provide the evaluation script of ST-RRED , I want to compare my method with fastdvdnet using ST-RRED. Thanks.

[m-tassano]

Hey,

Sharing the whole script wouldn't make sense because it does many other things, but the computation of ST-RRED is basically what I copy below. Then, I average the scores of all the sequences in the testset per sigma.

from skvideo.measure import viideo_score, strred
from skvideo.io import vread
...

           # open ref seq
            try:
                refseq = vread(seq_ref_path, num_frames=nframes, outputdict={"-pix_fmt": "gray"})
            except:
                refseq = None
                
            # iterate over the list of sigmas and algorithms
            for sigma, algo in product(sigmaL, algoD.keys()):

                # Load denoised sequence only if there is no previously computed psnr
                if need_to_load_seq(psnr_list, algo, sigma, tsetname, subdir):
                    # Denoised: base_den_dir / algo / sigma / testset / seq
                    seq_den_path = os.path.join(base_den_dir, algo, tsetname, sigma, subdir, subdir+'.mp4')

                    # load denoised sequence for comparison
                    try:
                        denseq = vread(seq_den_path, num_frames=nframes, outputdict={"-pix_fmt": "gray"})
                    except:
                        denseq = None

                    # compute only if value had not been precomputed  
                    if (denseq is not None) and (refseq is not None):
                        
                        # compute psnr
                        psnr = psnrfunc(refseq, denseq).mean()
                        _, strred_fr, strred_rr = strred(refseq, denseq)